<< 30-08-2023 >>
00:40:17*ineptitude joined #nim
00:47:51FromDiscord<.maverk> is there any pdf version of the documentation ?
00:56:42*ineptitude quit (Quit: Leaving.)
00:59:13*ineptitude_ joined #nim
01:05:31*ineptitude_ quit (Quit: Leaving)
01:07:35*ineptitude joined #nim
01:09:37FromDiscord<Elegantbeef> Araq's answer is "print as pdf"
01:14:13*ineptitude quit (Quit: client terminated!)
01:19:48*rockcavera joined #nim
01:30:00*marcus quit (Remote host closed the connection)
01:31:48*marcus joined #nim
01:37:37FromDiscord<rainbowasteroids> sounds about right
02:44:06FromDiscord<sOkam! 🫐> is there a way to pass `attribute((someflag=val))` to the C side when exporting a function from Nim?
02:45:50FromDiscord<Elegantbeef> codegendecl
02:47:15FromDiscord<pk.kyle> do i use asyncnet for chronos or is there something inbuild in chronos?
02:47:28FromDiscord<Elegantbeef> Chronos has it's own http client
02:47:35FromDiscord<sOkam! 🫐> In reply to @Elegantbeef "codegendecl": that is SO useful for interop lol. ty!
02:47:38FromDiscord<pk.kyle> In reply to @Elegantbeef "Chronos has it's own": great!
02:50:02FromDiscord<Elegantbeef> It also uses bearssl so it's easy to ship
03:00:16FromDiscord<pk.kyle> how do i create a socket and send/recv messages?
03:01:17FromDiscord<Elegantbeef> Using?
03:01:57FromDiscord<pk.kyle> chronos
03:05:09FromDiscord<Elegantbeef> https://status-im.github.io/nim-chronos/docs/chronos/transports/datagram.html seems this module
03:05:30FromDiscord<pk.kyle> thx!
03:05:38FromDiscord<Elegantbeef> That's for udp atleast
03:06:49FromDiscord<pk.kyle> yeah udp is good for me
03:26:57FromDiscord<sOkam! 🫐> Is there a way to define the names for `import thing` programatically in some way, from a const earlier in the same file?
03:26:58FromDiscord<thedistantforest> sent a code paste, see https://play.nim-lang.org/#ix=4ESy
03:27:27FromDiscord<Elegantbeef> Use a macro to eat your code
03:27:56FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4ESz
03:28:25FromDiscord<sOkam! 🫐> i was thinking of something simpler than a macro, but i guess that could work 😔
03:28:30FromDiscord<Elegantbeef> Since you looke to be writing an ECS some inspiration can be got from https://github.com/beef331/nimtrest/blob/master/yeacs.nim
03:28:35FromDiscord<Elegantbeef> I mean you want to generate code
03:28:44FromDiscord<Elegantbeef> To generate code with an input you have to use macros
03:28:48FromDiscord<thedistantforest> In reply to @Elegantbeef "Since you looke to": ok thanks!
03:28:54FromDiscord<sOkam! 🫐> yeah, can see that
03:29:40FromDiscord<Elegantbeef> It's \~500 loc, but it has a bunch ideas that can be taken, it does not use explicit ids for instance, but instead uses Nim's RTTI pointer to distinguish types
03:29:50FromDiscord<Elegantbeef> This means you can technically use any type you want
03:30:04FromDiscord<Elegantbeef> Be it an integer, float, bool, string, ...
03:30:26FromDiscord<Elegantbeef> Though this does mean it does not have persistent IDs
03:31:43FromDiscord<Chronos [She/Her]> In reply to @Elegantbeef "Chronos has it's own": I do? 😛
03:32:29FromDiscord<Chronos [She/Her]> Hm... Why can't all libs that use asyncdispatch, work with chronos instead? Think someone talked about it
03:33:11FromDiscord<Elegantbeef> Cause the implementations are different and no one thought to make async procedures generics
03:34:40FromDiscord<Chronos [She/Her]> Ah rip
03:34:57FromDiscord<Elegantbeef> @thedistantforest You're planning on making a proper ECS and not a Unity style ECS right?
03:35:27FromDiscord<Chronos [She/Her]> Is that hard to do?
03:35:46FromDiscord<Elegantbeef> Is what hard to do?
03:37:29FromDiscord<thedistantforest> In reply to @Elegantbeef "<@262410914261762050> You're planning on": Yeap, well actually I'm trying to port an existing ecs solution that I'm quite comfortable with
03:37:31FromDiscord<thedistantforest> https://gitea.moonside.games/MoonsideGames/MoonTools.ECS
03:38:14FromDiscord<thedistantforest> When i was messing around with lua, it was pretty easy to port there, but was running into to performance issues (probably because I made a quick and dirty port)
03:40:19FromDiscord<Chronos [She/Her]> In reply to @Elegantbeef "Is what hard to": ...i actually don't know tbh, making async procedures generic ig?
03:40:39FromDiscord<thedistantforest> In reply to @thedistantforest "Yeap, well actually I'm": Actually what I was trying to do is the store the type of a component into a hash map like what he has there. But I had no idea how to do that in nim so I settled with using strings 😌
03:40:39FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4ESA
03:40:43FromDiscord<Elegantbeef> Or just generating the id from the type name
03:40:54FromDiscord<Elegantbeef> You cannot store types at runtime in Nim
03:41:03FromDiscord<Elegantbeef> Well you kinda can, but you cannot really
03:41:45FromDiscord<thedistantforest> ok cool, so using string ids is probably an ok first pass solution..
03:42:10FromDiscord<Elegantbeef> The lasiest is just using a `pointer` as it's automatically generated
03:42:28FromDiscord<Elegantbeef> `getTypeInfo` is already implemented and it's a guaranteed unique ID
03:42:52FromDiscord<thedistantforest> interesting, ok
03:43:00FromDiscord<thedistantforest> i
03:43:01FromDiscord<thedistantforest> i'
03:43:06FromDiscord<thedistantforest> i'll check it out
03:43:23FromDiscord<Elegantbeef> You also should not store ID on types as it increase the size of the object
03:43:36FromDiscord<Chronos [She/Her]> In reply to @chronos.vitaqua "...i actually don't know": Would concepts be a useful thing here?
03:43:43FromDiscord<Elegantbeef> Probably
03:44:07FromDiscord<Chronos [She/Her]> Wonder why more effort isn't put into this then, oh well
03:44:34FromDiscord<Elegantbeef> It's a breaking change
03:44:35FromDiscord<thedistantforest> In reply to @Elegantbeef "You also should not": yeah makes sense
03:45:18FromDiscord<Chronos [She/Her]> In reply to @Elegantbeef "It's a breaking change": True but it'd also lead to better QoL I'd imagine
03:45:55FromDiscord<Chronos [She/Her]> More libs available for either async impl
03:46:34FromDiscord<Elegantbeef> Ah the bigger issue now that I think about it is that async is a macro and cannot be used with a generic that way
03:47:04FromDiscord<Chronos [She/Her]> Ah
03:47:15FromDiscord<Chronos [She/Her]> I do not understand that at all 😎
03:47:27FromDiscord<Elegantbeef> Macros evaluate at declaration not instantiation
03:47:41FromDiscord<Elegantbeef> As such you cannot have your specific async macro run that generates your specific code
03:49:14FromDiscord<Chronos [She/Her]> I still don't get it-
03:49:43FromDiscord<Elegantbeef> The async macro for each library is different, you need to use a different one when you use a different runtime
03:50:00FromDiscord<Chronos [She/Her]> Okay that makes sense, ty-
03:50:49FromDiscord<Chronos [She/Her]> Smooth-brain momennt
04:52:03*cedb quit (Ping timeout: 255 seconds)
04:53:47*ced1 joined #nim
04:56:16*ced1 is now known as cedb
05:28:19FromDiscord<m4ul3r> Is it possible to use `varargs[untyped]` in a proc?
05:28:28FromDiscord<Elegantbeef> Nope
05:28:50FromDiscord<m4ul3r> hmm, I'
05:29:40FromDiscord<m4ul3r> (edit) "I'" => "I'm trying to write one function that handles different amount of args and types, with the logic of handling it in the one func.↵I know I can overload the function, but I'd rather write one and handle what I'm passing in"
05:30:18FromDiscord<Elegantbeef> Make a macro that takes in varargs[typed] then do everything in there
05:31:28FromDiscord<m4ul3r> I'm not sure if I can do that since I'm using inline asm
05:37:21FromDiscord<Elegantbeef> Of course you can
05:43:51FromDiscord<m4ul3r> When I try, i get Error: cannot generate VM code for asm "
05:44:01FromDiscord<Elegantbeef> Well what are you doing
05:44:20FromDiscord<Elegantbeef> You need to return asm not write asm inside the macro
05:45:49FromDiscord<jviega> Hey beef, when trying to port stuff to Nim 2.0, I'm not comprehending what to do about this one:
05:45:55FromDiscord<m4ul3r> I'm passing the args and handling with asm
05:45:57FromDiscord<jviega> sent a code paste, see https://play.nim-lang.org/#ix=4EST
05:46:10FromDiscord<jviega> Errors on the MkSeq case
05:46:15FromDiscord<Elegantbeef> Well you're doing it wrong, if you provided what you're doing I could say more
05:46:20FromDiscord<jviega> `Users/viega/.choosenim/toolchains/nim-2.0.0/lib/system/comparisons.nim(309, 6) Error: '==' can have side effects`
05:46:29FromDiscord<jviega> I'm working on it 🙂
05:46:41FromDiscord<odexine> In reply to @jviega "`Users/viega/.choosenim/toolchains/nim-2.0.0/lib/sy": Oh I hit this issue too some time ago I feel
05:46:43FromDiscord<jviega> It doesn't have side effects, I went and looked at the code in comparisons.nim
05:46:44FromDiscord<Elegantbeef> Talking to the other person there jvg
05:46:55FromDiscord<odexine> Someone naively swapped to func I feel
05:47:15FromDiscord<Elegantbeef> Yea there is a generic somewhere marked func I think
05:47:16FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4ESU
05:47:28FromDiscord<jviega> So what's the workaround?
05:47:53FromDiscord<odexine> I think it doesn’t happen on devel
05:47:58FromDiscord<Elegantbeef> Define your own `==` comparison for that type
05:48:10FromDiscord<m4ul3r> sent a code paste, see https://play.nim-lang.org/#ix=4ESV
05:48:17FromDiscord<Elegantbeef> Exactly like i said
05:48:19FromDiscord<jviega> Chronos, that's just a matter of preference. I think people generally have an easier time with more explicit code.
05:48:20FromDiscord<Elegantbeef> You're not using a macro as a macro
05:48:27FromDiscord<Elegantbeef> Use the macro as a macro and problem solved
05:48:46FromDiscord<Chronos [She/Her]> In reply to @jviega "Chronos, that's just a": That's fair, tbf I'd do the same really
05:48:58FromDiscord<Elegantbeef> A macro has to return AST and does not have a existent function
05:49:16FromDiscord<jviega> It doesn't have any impact to speed, just to clarity
05:49:31FromDiscord<Chronos [She/Her]> Yeah I'm aware aha
05:49:33FromDiscord<Elegantbeef> You can do `proc doThing[T](arg: varargs[T])` if you do not need heterogenous values
05:50:07FromDiscord<m4ul3r> I had tried that, but my problem is heterogenous types
05:50:33FromDiscord<Elegantbeef> Well then like i said make a macro that does what your asm did
05:51:04FromDiscord<Elegantbeef> No you cannot just paste in your asm in a macro you need to construct a `nnkAsmStmt` with the coresponding code inside of it
05:51:20FromDiscord<m4ul3r> thanks for the advice, i'll have to figure out more about macros. They're pretty confusing to me
05:52:12FromDiscord<Elegantbeef> Ah yea so only some default `==`s have nosideeffect in 2.0
05:52:27FromDiscord<Elegantbeef> `object` is not a func, nor is `openArray`
05:52:39FromDiscord<Elegantbeef> but `seq[T]` is so if there is a side effect inside an `==` it'll error
05:53:08FromDiscord<Elegantbeef> You might be able to wrap that body with `{.cast(noSideEffect).}:` albeit ugly it'd get around it
05:54:07FromDiscord<Elegantbeef> Don't have the error stack in front of me so uncertain where the error is actually bubbling from 😄
05:54:18FromDiscord<jviega> I think that's on the list I tried, but I hadn't tried my on comparison
05:54:24FromDiscord<jviega> I'd tried comparing the pointers
05:54:47FromDiscord<jviega> Now I'm just recursively decomposing really. It's lame but ok
05:55:18FromDiscord<Elegantbeef> Nah you do not compare seq pointers
05:55:18FromDiscord<Elegantbeef> They're value types
05:55:19FromDiscord<Elegantbeef> Comparison is O(N)
05:55:31FromDiscord<Elegantbeef> This does matter since of course since you might hash this type, and if your `==` and `hash` do not align, you get incorrect behaviour
05:56:02FromDiscord<jviega> Well, it depended on the implementation
05:56:08FromDiscord<jviega> Right
05:56:25FromDiscord<Phil> In reply to @m4ul3r "thanks for the advice,": Maybe it helps you in terms of understanding as it was an aha moment for me:↵Every piece of Nim Syntax is a fixed pattern of nimnodes.↵All you need to do is figure out what Nim Syntax you want to replicate, look at its pattern e.g. with dumpastgen and have a macro return a nimnode containing the pattern of that syntax
05:56:27FromDiscord<jviega> I'm all good now, just manually comparing. Don't know why I didn't do that right away
05:56:41FromDiscord<Elegantbeef> Wait what was the fix?
05:57:04FromDiscord<Elegantbeef> Stop suggesting dumpastgen in the year 2023, thank
05:57:07FromDiscord<Elegantbeef> Use `treeRepr` or `dumpTree`
05:57:17FromDiscord<jviega> sent a code paste, see https://play.nim-lang.org/#ix=4ESY
05:57:28FromDiscord<Phil> I keep forgetting the tree version, also: nevaaa
05:57:34FromDiscord<Elegantbeef> Right but that's just `box1.c.s == box2.c.s` 😄
05:57:44FromDiscord<jviega> Yeah but that statement fails
05:57:51FromDiscord<jviega> Because nim thinks it side-effects??
05:57:53FromDiscord<Phil> It's imo way better for getting a grasp of the basics
05:58:03FromDiscord<Elegantbeef> Right and what is the type of `box1.c.s`?
05:58:07FromDiscord<jviega> seq[Box]
05:58:24FromDiscord<Elegantbeef> Right so force your proc to have no side effect
05:58:36FromDiscord<Elegantbeef> So you do not have to reimplement seq comparisons
05:58:57FromDiscord<jviega> I'll try again, one sec
05:59:07FromDiscord<Elegantbeef> alternatively do `box1.c.s.toOpenArray(0, box1.high) == box2.c.s.toOpenArray(0, box2.high)`
05:59:16FromDiscord<Elegantbeef> Since that is not a pure proc
05:59:59FromDiscord<Elegantbeef> Yea Phil I do not like mentioning it cause it inevitably ends with someone copying the code and using it instead of using `genast` or `quote do`
06:00:06FromDiscord<jviega> Where's the pragma go in that case? Before the colon I thought??
06:00:27FromDiscord<Elegantbeef> If you're casting the body it's a body level pragma
06:00:33FromDiscord<Elegantbeef> If you're annotating the proc it's before the `=`
06:01:15FromDiscord<Elegantbeef> sent a code paste, see https://play.nim-lang.org/#ix=4ET0
06:01:15FromDiscord<Elegantbeef> If the latter doesnt work, the former will
06:01:23FromDiscord<Phil> I regard dumpastgen in macros like I regard vanilla ja: necessary to learn to get the fundamentals, before you it makes sense to use "higher level tooling"
06:01:37FromDiscord<Phil> (edit) "ja:" => "js:"
06:01:47FromDiscord<Elegantbeef> bleh learning to look at treerepr gets the same thing as astgenrepr ime, but meh
06:01:54FromDiscord<Phil> (edit) removed "you"
06:01:59FromDiscord<Elegantbeef> Not going to argue
06:02:59FromDiscord<jviega> The statement level one was the one I'd tried before, that doesn't work (doesn't complain about the bragma but still gives the error), but the later does, believe it or not.
06:03:07FromDiscord<jviega> So thanks again
06:03:15FromDiscord<Elegantbeef> Likely just an issue with Nim's func tracking
06:03:25FromDiscord<Elegantbeef> you can of course just declare it using the `func` keyword instead
06:05:02FromDiscord<jviega> If it seems to think the base comparison operator for seqs side effects, I'd have expected that to fail??
06:05:29FromDiscord<jviega> Anyway, this is good enough, thanks very much
06:05:38FromDiscord<Elegantbeef> Ah actually It's a known issue with the effects
06:05:57FromDiscord<Elegantbeef> Since this `==` is recursive and `seq`s require no side effect, it does not mark it func
06:06:16FromDiscord<Elegantbeef> Cause it doesnt know it's func cause at the point of the seq invokation `==` it's not done analysing the procedure
06:06:29FromDiscord<Elegantbeef> It's explicitly needed due to the hidden recursion
06:07:01FromDiscord<Elegantbeef> That's also why the cast does not work
06:07:11FromDiscord<jviega> Got it
06:07:33FromDiscord<Elegantbeef> As you cast only the body, when the `seq[Box]`'s`==` is instantiated `Box`'s `==` is marked `proc`
06:08:50FromDiscord<Elegantbeef> In hindsight it's a pretty reasonable issue, albeit confusing to find
06:09:48FromDiscord<Elegantbeef> To make it less error prone generic instantiations could be delayed, but that means that all bodies that were not semantically checked would need to be checked after instantiation, and if those have instantiations, they'd have to be delayed.....
06:10:22FromDiscord<Elegantbeef> Very recursive solution vs. `func` 😛
06:20:42*rockcavera quit (Remote host closed the connection)
06:20:43*ntat joined #nim
06:32:08*advesperacit joined #nim
06:34:11FromDiscord<tommy_plug> sent a long message, see http://ix.io/4ET3
06:42:48FromDiscord<odexine> oh here too huh
06:42:50FromDiscord<odexine> <@&371760044473319454>
06:42:52FromDiscord<arathanis> @technorazor
06:42:58FromDiscord<arathanis> you beat me to it
06:43:29FromDiscord<arathanis> also i pinged the wrong person, sorry anosmicoder
06:43:57FromDiscord<pmunch> Hmm, strange, I had already deleted it on the Matrix side
06:44:25FromDiscord<odexine> In reply to @arathanis "also i pinged the": lol
06:47:05FromDiscord<Elegantbeef> Arath calls emergency services then goes "operator, operator I need... uhhh"
06:47:32FromDiscord<arathanis> > operator, operator, i'd like to order a pizza
06:48:06FromDiscord<arathanis> the tab complete failed me, i always forget discord keeps role mentions damn near the bottom of the list. it would really rather you ping an individual lol
06:48:19FromDiscord<arathanis> https://media.discordapp.net/attachments/371759389889003532/1146335572173017128/image.png
06:55:49FromDiscord<toma400> sent a code paste, see https://play.nim-lang.org/#ix=4ETa
06:56:32FromDiscord<toma400> (edit) "https://play.nim-lang.org/#ix=4ETa" => "https://play.nim-lang.org/#ix=4ETb"
06:56:43FromDiscord<Elegantbeef> try `[F, S](p: pair[F, S])`
06:56:47FromDiscord<toma400> (edit) "https://play.nim-lang.org/#ix=4ETb" => "https://play.nim-lang.org/#ix=4ETc"
06:56:51FromDiscord<enthus1ast> Elegantbeef\: whats so bad about dumpAstGen? It produces stuff i can copy and paste into a macro
06:56:56FromDiscord<Elegantbeef> Likely a bug with the implicit generic
06:57:03FromDiscord<Elegantbeef> It's unmaintainable
06:57:23FromDiscord<enthus1ast> whats the lazy ass alternative then?
06:57:31FromDiscord<Elegantbeef> `genast`/`quotedo`
06:57:48FromDiscord<Elegantbeef> use genast though
06:59:37FromDiscord<toma400> In reply to @Elegantbeef "Likely a bug with": Would you be so kind to explain shortly what this means in simpler terms?↵I just got into generics, so it fits that my use of it is silly/wrong 😅
06:59:54FromDiscord<Elegantbeef> `pair` is an implicit generic
07:01:01*PMunch joined #nim
07:01:13FromDiscord<enthus1ast> it could be because i just woke up but i fail to understand beef or, how this is an alternative to "get the ast that produces this nim code in a copy pastable form"
07:01:43FromDiscord<Elegantbeef> It completely replaces that step
07:02:35FromDiscord<Elegantbeef> You do not need to have ast procedures in your code, you just have the genast in your macro
07:18:16FromDiscord<Elegantbeef> As an example https://github.com/beef331/micros/blob/6791af3cacfdf1c098dc404376b09865c1a1d2cc/src/micros/definitions/objectdefs.nim#L55-L68 this is much more readable than the astgenrepr version
07:33:08*ntat quit (Quit: leaving)
07:38:10FromDiscord<System64 ~ Flandre Scarlet> https://github.com/mratsim/constantine/blob/1cb6c3d/constantine/math/arithmetic/assembly/limbs_asm_mul_x86_adx_bmi2.nim#L90-L111↵Nim supports some Assembly?
08:19:22termeryou can inline assembly just like in C
08:20:12termerIt looks like that code might be using macros to do it though
09:05:49*calebjohn quit (Ping timeout: 250 seconds)
09:06:15*calebjohn joined #nim
10:03:42*azimut quit (Ping timeout: 246 seconds)
10:05:35*azimut joined #nim
10:16:40FromDiscord<System64 ~ Flandre Scarlet> Wow, pretty powerful!
10:17:19FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4ETR
11:34:07*jmdaemon quit (Ping timeout: 245 seconds)
11:36:21PMunchHmm, what's the best bignum library out there nowadays?
11:36:54PMunchI want to overhaul my calculator thing, and floats don't really cut it..
11:47:57PMunchHmm, can't really find anything with support for trig functions.. I guess I could convert to float for those
11:49:49FromDiscord<odexine> i believe its especially complex to implement those for bignums
11:49:57FromDiscord<odexine> or rather true-decimals
11:50:21PMunchI did find MAPM, which seems to be exactly what I need, but it seems to have died.. I pretty much only find archive.org links like this one: https://web.archive.org/web/20161218111256/http://www.tc.umn.edu/~ringx004/article-main.html
11:51:12PMunchFair, I believe the trig functions are often table-lookup + interpolation anyways, so precision won't need to be great
11:51:34FromDiscord<odexine> are you looking for arbitrary precision or just better than floats?
11:54:29PMunchWell really just better than float. But since I'm doing a complete rewrite anyways I figured I could go all the way
12:02:21FromDiscord<t0ngub1n> sent a code paste, see https://play.nim-lang.org/#ix=4EUe
12:12:44FromDiscord<mratsim> In reply to @PMunch "Well really just better": mpdecimal.
12:12:57FromDiscord<mratsim> The state of bignum in Nim is pretty sad at the moment
12:13:08FromDiscord<mratsim> I know how to fix it but I have no time.
12:14:25FromDiscord<mratsim> Ah there is a full native conversion by JohnAD that is WIP
12:14:34FromDiscord<mratsim> But I was thinking of wrapping this: https://github.com/status-im/nim-decimal/blob/master/LICENSE-BSD-mpdecimal
12:14:42FromDiscord<mratsim> (edit) "https://github.com/status-im/nim-decimal/blob/master/LICENSE-BSD-mpdecimal" => "https://github.com/status-im/nim-decimal"
12:15:18FromDiscord<System64 ~ Flandre Scarlet> Where can I know there is a deep copy going on? https://media.discordapp.net/attachments/371759389889003532/1146417860156076102/message.txt
12:16:14PMunchYeah I looked at mpdecimal, but it didn't have trig functions (which again isn't a big problem, but MAPM supports them)
12:16:26PMunchPlaying around with MAPM through Futhark now, seems to work pretty well
12:17:01PMuncht0ngub1n, did you run finish.exe after installing Nim?
12:17:04FromDiscord<mratsim> In reply to @PMunch "Hmm, can't really find": You can search for Pade approximant or Rehmer (?) polynomials for approximating trigonometric function with arbitrary precision
12:17:26PMunchmratsim, sure but that's more work, and I risk getting it wrong
12:17:35FromDiscord<mratsim> All the SIMD libraries use that or table based interpolation
12:18:28FromDiscord<odexine> oh ive recently learned about pade approx
12:18:29FromDiscord<odexine> p cool
12:34:39FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EUk
12:34:50FromDiscord<System64 ~ Flandre Scarlet> (PlayerORM are ref objects)
12:37:46FromDiscord<System64 ~ Flandre Scarlet> I'm afraid it leads to unexpected results, so bugs
13:37:02*cedb quit (Ping timeout: 245 seconds)
13:39:07*ced2 joined #nim
13:45:10*advesperacit quit ()
13:48:43FromDiscord<t0ngub1n> In reply to @PMunch "<@789234515268599818>, did you run": yes
13:49:09FromDiscord<t0ngub1n> (edit) "yes" => "yes, i did"
13:52:54PMunchThat should've installed all dependencies. Maybe your AV deleted some of them?
13:53:36*rockcavera joined #nim
13:58:48FromDiscord<t0ngub1n> no, it did not
13:58:54FromDiscord<t0ngub1n> the error is clearly in ld
13:59:00FromDiscord<t0ngub1n> or libraries
13:59:16FromDiscord<t0ngub1n> clang also says that "thread-local storage is not supported for the current target"
14:00:00FromDiscord<t0ngub1n> so maybe that
14:00:27PMunchCould you run gcc --version and clang --version? Just curious
14:00:44PMunchAnd does this happen with all Nim code? Even just a simple hello world program?
14:00:46*PMunch quit (Quit: Leaving)
14:01:03FromDiscord<t0ngub1n> yes, even with hello world
14:02:17FromDiscord<t0ngub1n> sent a code paste, see https://play.nim-lang.org/#ix=4EUI
14:08:24FromDiscord<t0ngub1n> has anyone encountered such a problem b4?
14:22:44FromDiscord<burnysc2> sent a code paste, see https://play.nim-lang.org/#ix=4EUS
14:31:21FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EUV
14:43:37*ced2 is now known as cedb
14:47:34FromDiscord<Chronos [She/Her]> @burnysc2 I have a solution for ya, it uses generics! Let me send it rq
14:48:30FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4EV2
14:48:57FromDiscord<burnysc2> In reply to @chronos.vitaqua "<@121948437796093952> I have a": Thanks but I've already solved it by replacing `proc step` with `method step`, that worked for me
14:49:10FromDiscord<burnysc2> Was a hard find in the manual
14:49:12FromDiscord<Chronos [She/Her]> By using generics, we don't lose the original type data, letting `step` from b.nim run, if that makes sense?
14:49:35FromDiscord<Chronos [She/Her]> Ah fair aha, if you wanna avoid dynamic dispatch this may help anyway, wish ya luck!
14:50:33FromDiscord<burnysc2> I think your solution may come in handy in the future for me, so thank you!↵Here `T` has be or inherits from the `type Bot` ?
14:51:53FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EV4
15:04:36FromDiscord<Chronos [She/Her]> In reply to @burnysc2 "I think your solution": Of course! Also T has to inherit from `Bot`, yeah
15:06:37FromDiscord<sOkam! 🫐> If I build from a host linux system, for a windows system with cross-compilation...↵does the value of `hostOS` change to windows?
15:06:51FromDiscord<sOkam! 🫐> (edit) "windows?" => "windows in the resulting binary?"
15:08:41FromDiscord<griffith1deadly> sent a code paste, see https://play.nim-lang.org/#ix=4EVe
15:11:37FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "except CatchableError as error:": https://media.discordapp.net/attachments/371759389889003532/1146462234625720381/image.png
15:12:08FromDiscord<griffith1deadly> echo error[]
15:12:25FromDiscord<burnysc2> sent a code paste, see https://play.nim-lang.org/#ix=4EVf
15:12:57FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://paste.rs/jMcbM
15:13:08FromDiscord<sOkam! 🫐> sent a code paste, see https://play.nim-lang.org/#ix=4EVh
15:13:27FromDiscord<jviega> From where?
15:13:33FromDiscord<jviega> Depends a great deal
15:13:41FromDiscord<jviega> Because nimscript runs on rosetta
15:13:48FromDiscord<jviega> So it'll always answer amd64
15:13:52FromDiscord<sOkam! 🫐> cross-compiling
15:14:06FromDiscord<sOkam! 🫐> (edit) "cross-compiling ... " added "a binary"
15:14:11FromDiscord<odexine> sent a code paste, see https://play.nim-lang.org/#ix=4EVi
15:14:29FromDiscord<jviega> Here's what I use in config.nims:
15:14:35FromDiscord<jviega> sent a code paste, see https://play.nim-lang.org/#ix=4EVj
15:14:41FromDiscord<griffith1deadly> sent a code paste, see https://play.nim-lang.org/#ix=4EVk
15:15:52FromDiscord<sOkam! 🫐> sent a code paste, see https://play.nim-lang.org/#ix=4EVl
15:16:28FromDiscord<jviega> Correct, as far as nimscript is concerned, it's running on x86 any time it's on a mac as far as I can tell. It's getting the answer it gets from the machine because it's running in a Rosetta VM
15:16:31FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "check your variables for": You mean my code string?
15:16:58FromDiscord<sOkam! 🫐> In reply to @jviega "Correct, as far as": that saves me a lot of hours of headaches, ty so much man ❤️
15:17:04FromDiscord<jviega> My approach there always detects the underlying platform. It doesn't try to use the underlying compiler, since you'd always get the one in your VM
15:17:11FromDiscord<griffith1deadly> In reply to @sys64 "You mean my code": no, all variables for accessing db↵looks like someone get null
15:17:27FromDiscord<jviega> Yeah, no worries
15:18:08FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "no, all variables for": I'm using a connection pool so it shouldn't be null
15:18:17FromDiscord<jviega> The targetStr there is as you would pass to the compiler/linker if you need to pass it for anything (usually linking 🙂
15:19:16FromDiscord<jviega> From within your actual binaries, the testing you did above is right.
15:19:30FromDiscord<griffith1deadly> In reply to @sys64 "I'm using a connection": you get null in select statement
15:19:32FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "no, all variables for": My GameORM object is not null
15:20:32FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EVn
15:22:23FromDiscord<griffith1deadly> sent a code paste, see https://play.nim-lang.org/#ix=4EVo
15:23:17FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "try echo in withDbConn": So after con.select?
15:23:22FromDiscord<griffith1deadly> yes
15:24:53FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "yes": Not displayed
15:25:12FromDiscord<griffith1deadly> try before
15:26:10*deadmarshal_ quit (Quit: IRCNow and Forever!)
15:26:20FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "try before": displayed
15:26:31FromDiscord<griffith1deadly> In reply to @sys64 "displayed": not nil?
15:26:41FromDiscord<sOkam! 🫐> sent a code paste, see https://play.nim-lang.org/#ix=4EVq
15:27:00FromDiscord<sOkam! 🫐> I'm not compiling on mac, don't even have access to one, so I cannot afford to do that type of thing
15:27:55FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "not nil?": Not nil
15:29:39FromDiscord<griffith1deadly> In reply to @sys64 "Not nil": then error in nil obj variables↵https://github.com/moigagoo/norm/blob/develop/src/norm/model.nim#L108
15:30:05FromDiscord<sOkam! 🫐> sent a code paste, see https://paste.rs/oBJtX
15:30:26FromDiscord<Chronos [She/Her]> In reply to @burnysc2 "I got your version": Ah that's not how you're supposed to use the generic step-
15:30:40FromDiscord<griffith1deadly> (edit) "variables↵https://github.com/moigagoo/norm/blob/develop/src/norm/model.nim#L108" => "variables↵https://github.com/moigagoo/norm/blob/develop/src/norm/model.nim#L106"
15:31:04FromDiscord<Chronos [She/Her]> The generic is supposed to call the underlying proc (so `proc step(bot: CustomBotVariant)`)
15:31:16FromDiscord<Prestige> Is there a way to have an object variant where a property exists for 2 different "kinds" but not all of them?
15:31:32FromDiscord<Chronos [She/Her]> The point of the generic is that it's not going to follow one specific code path-
15:31:54FromDiscord<terrygillis> Hi is there anyway to get a mutable reference from a seq?
15:31:55FromDiscord<Chronos [She/Her]> In reply to @burnysc2 "I got your version": The code I sent should replace `a.nim` completely aha-
15:31:58FromDiscord<terrygillis> By that I mean
15:32:30FromDiscord<terrygillis> a function for seq that returns a ‘var T’ for element in seq
15:32:45FromDiscord<terrygillis> [] returns a lent
15:32:53FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4EVs
15:32:57FromDiscord<terrygillis> which is not mutable
15:33:17FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "then error in nil": FINALLY PROBLEM SOLVED !↵Child objects must not be nil
15:33:53FromDiscord<Chronos [She/Her]> In reply to @terrygillis "which is not mutable": Does `mitems` exist for sequences?
15:34:16FromDiscord<Chronos [She/Her]> In reply to @chronos.vitaqua "Ah that's not how": @burnysc2
15:34:53FromDiscord<burnysc2> In reply to @chronos.vitaqua "The code I sent": Ohh, I will try that in a sec
15:35:38FromDiscord<Chronos [She/Her]> Good luck ^^
15:36:13FromDiscord<sOkam! 🫐> In reply to @chronos.vitaqua "What are you trying": allowing to compile for both mac.x86 and mac.arm
15:38:10FromDiscord<terrygillis> In reply to @chronos.vitaqua "Does `mitems` exist for": Btw can I get rich text for discord mobile, because currently I can’t put in new lines or code in the chat box on the mobile app.
15:38:23FromDiscord<Chronos [She/Her]> Ah wait... @terrygillis your issue is that the sequence itself does not seem to be mutable?
15:38:27*xet7 joined #nim
15:39:27FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4EVv
15:39:36*edr joined #nim
15:40:11FromDiscord<Chronos [She/Her]> sent a code paste, see https://paste.rs/McI2A
15:40:17FromDiscord<odexine> sent a code paste, see https://paste.rs/ZvjVW
15:40:27FromDiscord<Chronos [She/Her]> Oh did I misunderstand it? F
15:40:58FromDiscord<Chronos [She/Her]> In reply to @heysokam "allowing to compile for": What's the point of the command?
15:45:04FromDiscord<Prestige> Is there some way to do this? https://play.nim-lang.org/#ix=4EVw↵Note the duplicate `t` variable
15:45:27FromDiscord<jviega> @sOkam! 🫐 That command will exec a process, but Rosetta is a VM, so the command runs in the VM
15:45:42FromDiscord<jviega> And how it responds tells us what arch we're really on
15:46:01FromDiscord<jviega> It's the only easy way I know from within the VM to determine if you're in the VM or not.
15:47:55FromDiscord<enthus1ast> @Prestige\: no unfortunately not
15:48:27FromDiscord<enthus1ast> there are a few discussions about this though
15:49:00*deadmarshal_ joined #nim
15:49:57FromDiscord<sOkam! 🫐> In reply to @jviega "It's the only easy": but how do you do that from windows or linux?
15:50:02FromDiscord<sOkam! 🫐> i don't have a mac
15:50:28FromDiscord<jviega> huh? If you're compiling on windows for a mac??
15:50:37FromDiscord<sOkam! 🫐> hardcoding `--cpu:arm64` forces the value of `hostCPU` afaik
15:50:39FromDiscord<jviega> If you're doing that, build two binaries
15:50:54FromDiscord<sOkam! 🫐> In reply to @jviega "huh? If you're": yeah, from linux for all system
15:50:59FromDiscord<sOkam! 🫐> (edit) "for" => "for/to"
15:51:00FromDiscord<enthus1ast> @Prestige\: here is one https://github.com/nim-lang/RFCs/issues/368
15:51:13FromDiscord<jviega> Universal binaries on Mac are actually two binaries ZIP'd together basically.
15:51:37FromDiscord<sOkam! 🫐> yeah but i don't want to build two things and have to link them myself
15:51:49FromDiscord<sOkam! 🫐> this buildsystem has enough bs how it is
15:51:54FromDiscord<Prestige> Ty
15:52:43FromDiscord<jviega> Well then you have what you need for making it easy for people to build with the right libs and flags on a mac
15:52:58FromDiscord<jviega> hostOs and hostCPU should be good enough everywhere else
15:53:17FromDiscord<sOkam! 🫐> but im not telling mac people "hey, to play my game, first install nim and compile it and then figure out how to run it and everything"
15:53:29FromDiscord<jviega> Then you'd better build on a mac?
15:53:37FromDiscord<sOkam! 🫐> or just use zigcc?
15:53:52FromDiscord<sOkam! 🫐> zigcc builds for all systems, and it runs well. i tried it. but never gave support for arm
15:53:57FromDiscord<sOkam! 🫐> so i havent tested that part
15:53:59qwrif they compile, why not build for specific arch? i.e. not use the universal binary
15:54:12FromDiscord<sOkam! 🫐> In reply to @qwr "if they compile, why": yeah that's what im doing
15:54:15FromDiscord<jviega> You could cross-compile, but you're going to have linking issues everywhere if you don't statically link, so you need .a's for everything wherever you're compiling
15:54:34FromDiscord<sOkam! 🫐> (edit) "doing" => "doing↵the idea is finding which is running from the compiler so it can link the correct file/flags"
15:54:36FromDiscord<jviega> And cross-compiling is more likely to have issues
15:55:14FromDiscord<sOkam! 🫐> In reply to @jviega "You could cross-compile, but": is that true if im only depending on libc and that is figured out by zigcc?
15:55:14FromDiscord<jviega> So I don't know what you're still looking for. I gave you everything you need to know, and then it's just a matter of what's right for you in terms of where and how to build
15:55:19FromDiscord<sOkam! 🫐> the rest of my libs are all static
15:55:39FromDiscord<jviega> If they're static you need to link in the static versions for those platforms.
15:55:47FromDiscord<jviega> That's exactly what I'm saying
15:55:56FromDiscord<sOkam! 🫐> In reply to @jviega "So I don't know": i was just asking because what you sent is for building -from- a mac. and i don't have one. so was looking for alternative
15:56:15FromDiscord<sOkam! 🫐> In reply to @jviega "If they're static you": yeah i figure
15:56:15FromDiscord<jviega> For what???
15:56:21FromDiscord<sOkam! 🫐> you bring agood point though
15:56:31FromDiscord<jviega> You have someone pass a flag to tell you what arch to build for if you're cross-compiling
15:56:40FromDiscord<jviega> I don't get why that requires anything special
15:57:14FromDiscord<sOkam! 🫐> i was asking because the snippet does `staticExec` and that relies on the buildsystem running on a mac mandatory
15:57:46FromDiscord<jviega> That's because the snippit detects the underlying OS even though nimscript always runs in the vm
15:57:50FromDiscord<enthus1ast> @sOkam! 🫐\: i would create nimble tasks for all the architectures you wanna support
15:57:57FromDiscord<jviega> That's what it sounded like you were asking
15:57:59FromDiscord<enthus1ast> then just call the correct compiler \:)
15:58:09FromDiscord<jviega> If you want to build on a mac and get the right architecture, that's how you do it
15:58:16FromDiscord<jviega> No matter who is building
15:58:37FromDiscord<jviega> If you want to cross-compile, cross-compile
15:58:46FromDiscord<jviega> That's what the --arch flag is for
15:59:13FromDiscord<sOkam! 🫐> In reply to @enthus1ast "<@186489007247589376>\: i would create": thats not the issue. the problem is understanding which system is being built from the file, so i can conditionally import each correct set of flags↵we are talking 60-70 variants... so manually writing nimble tasks for each is out of question fully
15:59:48FromDiscord<sOkam! 🫐> kk ty jtv
15:59:53FromDiscord<enthus1ast> i've done cross cross compilation like this\: https://github.com/enthus1ast/nimDownloadGrist/blob/master/nimGristDownload.nimble
16:00:23FromDiscord<sOkam! 🫐> In reply to @enthus1ast "i've done cross cross": does `--cpu` change the value of `hostCPU`?
16:00:28FromDiscord<sOkam! 🫐> because that was my original question
16:00:45FromDiscord<sOkam! 🫐> I reworded it because it seem to not be understood
16:01:05FromDiscord<sOkam! 🫐> or nobody knew, one of those
16:01:48FromDiscord<jviega> No
16:01:54FromDiscord<jviega> It doesn't
16:02:10FromDiscord<jviega> hostCPU is all about what's the environment you're running in right now
16:02:18FromDiscord<jviega> Whether that's nimscript or a binary
16:04:05FromDiscord<enthus1ast> i once got kicked and banned from a macos channel because i asked how to create a macos vm, since then i could care less about macos \:)
16:05:02FromDiscord<jviega> Oh that's prob because Mac doesn't license it to run on 3rd party hardware and they go after people who seem to support that kind of thing
16:05:08FromDiscord<sOkam! 🫐> sent a code paste, see https://play.nim-lang.org/#ix=4EVH
16:07:09FromDiscord<sOkam! 🫐> clientvk + clientgl + clientDL + server + servergame + clientgame + ui + vulkan.so + opengl.so 2x (debug,release) 4 (mac1, mac2, lnx, win)↵its.... just something else
16:07:50FromDiscord<jviega> That snippit doesn't work in a config.nims or a nimble file on a mac anyway
16:08:03FromDiscord<sOkam! 🫐> win,mac,lnx i can solve no issues. but the arm-vs-x86 is what is giving me the trouble
16:08:11FromDiscord<jviega> Like I said, use the code I gave you if you need to detect the platform when building on a mac
16:08:15FromDiscord<sOkam! 🫐> In reply to @jviega "That snippit doesn't work": yeah i know, that's why i asked in the first place
16:08:29FromDiscord<sOkam! 🫐> In reply to @jviega "Like I said, use": yeah, i understand jtv, but i don't build on a mac
16:08:41FromDiscord<enthus1ast> when you cannot test on a mac, i see no reason to actually build for a mac to be honest
16:08:41FromDiscord<jviega> Otherwise, accept a flag to tell you what platform to build from!!!
16:08:52FromDiscord<sOkam! 🫐> and im a solodev, and don't have the luxury of affording a mac, not even a cheap one
16:08:56FromDiscord<jviega> How is that part hard?
16:09:19FromDiscord<sOkam! 🫐> In reply to @enthus1ast "when you cannot test": i can test on an x86 mac, from a user/friend. but not on arm
16:09:23FromDiscord<jviega> sent a code paste, see https://play.nim-lang.org/#ix=4EVI
16:09:45FromDiscord<sOkam! 🫐> In reply to @jviega "There's an example of": i don't build on a mac, and the snipped relies on it
16:09:46FromDiscord<jviega> That will then accept -d:arch=whatever
16:09:51FromDiscord<jviega> No it doesn't
16:09:56FromDiscord<sOkam! 🫐> i was asking if i understood the snippet correctly or not
16:10:14FromDiscord<jviega> The flag setup assumes a native build OR you can override
16:10:34FromDiscord<jviega> It's set up in such a way that it's only possible to override on a mac, because I'm not a huge fan of cross-compiling
16:10:51FromDiscord<jviega> But it's got everything you need to know in there
16:11:09FromDiscord<sOkam! 🫐> well i asked "does this rely on being on a mac? because i don't have one"... and then the conversation turned into this silly debate on why or how or etc. i was asking from the start if I understood or not
16:11:20FromDiscord<sOkam! 🫐> im that obtuse, so im confident to accept when I misread things
16:11:26FromDiscord<sOkam! 🫐> hence why i started with that
16:11:46FromDiscord<jviega> It's not a debate, it's me trying to understand what you really meant, and then try to get you to understand what you need to do, depending on your requirements
16:11:51FromDiscord<jviega> I don't care what you do 🙂
16:12:07FromDiscord<jviega> I don't even know what the debate would be there
16:12:24FromDiscord<sOkam! 🫐> In reply to @jviega "But it's got everything": kk. i saved the snipped in place, so will try to test it if I can lure anyone on macarm to try it out
16:16:01FromDiscord<terrygillis> In reply to @chronos.vitaqua "Ah wait... <@908363885667553300> your": nevermind I realize it’s technically impossible without touching unsafe code
16:16:07FromDiscord<enthus1ast> @sOkam! 🫐\: https://github.com/nim-lang/Nim/blob/2e4e2f8f5076b39e5976ba20f231f468b1b5052c/tests/compiler/tcmdlinecpuamd64.nim#L2
16:16:09FromDiscord<enthus1ast> it seems hostCPU IS set by --cpu
16:16:32FromDiscord<sOkam! 🫐> In reply to @enthus1ast "it seems hostCPU IS": ty! that is -so- helpful to know ✍️
16:26:37FromDiscord<jviega> Interesting, def good to know!
16:31:29*ntat joined #nim
16:32:33*azimut quit (Ping timeout: 246 seconds)
16:35:46*krux02 joined #nim
16:52:26FromDiscord<Chronos [She/Her]> In reply to @terrygillis "nevermind I realize it’s": Hm
16:53:35FromDiscord<Chronos [She/Her]> Unrelated, but should I make my event system library integrate with JS' event system stuff when compiled for the web?
17:11:09FromDiscord<Phil> sent a long message, see http://ix.io/4EW2
17:11:48FromDiscord<Phil> (edit) "http://ix.io/4EW2" => "http://ix.io/4EW3"
17:12:45*advesperacit joined #nim
17:16:50FromDiscord<Chronos [She/Her]> In reply to @isofruit "Argh I came back": Oh Phil, you exist! If you know by any chance, how would I store a sequence of bytes in a pg database? :P
17:16:55FromDiscord<Chronos [She/Her]> With Norm specifically
17:17:09FromDiscord<Chronos [She/Her]> Wanted to use it a while ago but there wasn't really a way for me to do that aha
17:17:28FromDiscord<Phil> Hmm note that you're limited by whatever variants dbType provides
17:17:50FromDiscord<Phil> Because it goes↵your type <--> dbType <--> Whatever the heck it is store in the Database as
17:18:07FromDiscord<Phil> If you can store that seq of bytes as a string or a list of numbers or whatever and convert it back, you're golden
17:18:51FromDiscord<Phil> Basically you need to act as if it's a custom datatype and then just do what is described here:↵https://norm.nim.town/customDatatypes.html
17:18:56FromDiscord<Phil> (edit) "store" => "stored"
17:19:37FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4EW5
17:20:12FromDiscord<Phil> Simply regard seq[byte] as a custom type.↵If it helps with the abstraction, just alias `type MyBytes = seq[byte]` or something and define the procs accordingly
17:20:13FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "Argh I came back": Ah yeah I can put default values
17:22:05FromDiscord<Chronos [She/Her]> In reply to @isofruit "If you can store": What if there's a null byte though hm...
17:22:21FromDiscord<Chronos [She/Her]> In reply to @isofruit "Simply regard seq[byte] as": Gucci, thanks Phil!
17:22:28FromDiscord<Phil> In reply to @chronos.vitaqua "What if there's a": Are those allowed?
17:22:36FromDiscord<Phil> If so, you could do seq[Option[byte]]
17:22:54FromDiscord<Phil> If you need to check "does this exist", assuming that a null-byte means "does not exist"
17:23:01FromDiscord<Phil> (edit) ""does" => ""this byte does"
17:23:58FromDiscord<Chronos [She/Her]> In reply to @isofruit "Are those allowed?": In Nim? Yeah
17:24:07FromDiscord<Chronos [She/Her]> Even in Nim strings, which iirc Postgres does not like
17:24:20FromDiscord<Phil> Assume I don't know what a null byte means for sure, i just assume all bits are value 0 and that expressed as an int just means 0
17:24:51FromDiscord<Chronos [She/Her]> Yeah
17:25:14FromDiscord<Chronos [She/Her]> Since normally they're obviously used as string terminators but
17:25:34FromDiscord<Phil> How performance conscious are you about this?
17:26:04FromDiscord<Phil> Like every cycle matters or "Eh, it's okay if it works, postgres is fast enough and other parts of my algo will slow me down more"?
17:26:08FromDiscord<Chronos [She/Her]> Not too much? They're going to be used for lookup mostly though so
17:26:36FromDiscord<Phil> In that case your worst case fallback can always be to store a list of ints as text
17:26:49FromDiscord<Chronos [She/Her]> That sounds horrible lol
17:26:58FromDiscord<Chronos [She/Her]> But ig it's the best thing to do
17:27:14FromDiscord<Phil> Eh, it's inefficient sure but I would only bother looking for faster scenarios if it were performance critical
17:27:27FromDiscord<Phil> Or if I noticed it being a bottleneck
17:27:38FromDiscord<Chronos [She/Her]> Tbf my ID format is 128 bits long, if I could format it as a UUID somehow... If pg supports UUIDs
17:27:56FromDiscord<Chronos [She/Her]> In reply to @isofruit "Eh, it's inefficient sure": Storage would be a pain over time considering nearly everything would have an ID
17:28:04FromDiscord<Chronos [She/Her]> Messages, users, events
17:28:08FromDiscord<Chronos [She/Her]> Channels, servers
17:32:12FromDiscord<Phil> In reply to @chronos.vitaqua "Channels, servers": An alternative would be https://www.postgresql.org/docs/9.1/arrays.html↵But that one you would need to see whether you can squeeze that into norm, no promises there.↵That seems like db-specific syntax that norm doesn't support and hasn't had in mind so far, in part because I think the lower-level db it uses to deal with that (lowdb) might also not be capable of handling it.
17:32:51FromDiscord<Phil> But maybe it is, it's more of a "not tested" kind of thing than a "I know this doesn't work" and it's not terribly unlikely that this stuff can just get converted correctly
17:32:53FromDiscord<Chronos [She/Her]> Makes sense... Seems painful though
17:33:09FromDiscord<Chronos [She/Her]> I wonder if I'd be better off formatting my ID as a UUID
17:33:36FromDiscord<Phil> I mean, first things first, what is that seq[byte] actually that it's an id for you?
17:33:37FromDiscord<Chronos [She/Her]> UUIDs are supported in Postgres and Sqlite so
17:35:16FromDiscord<Chronos [She/Her]> In reply to @isofruit "I mean, first things": Contains a signed 48 bit timestamp, a unique 40 bit value, a 16 bit unique generator number, a 16 bit sequence (for the number of the ID that's been generated at the current time), and an 8 bit random value
17:36:51FromDiscord<Phil> What in the everloving...
17:37:11FromDiscord<Chronos [She/Her]> Aha
17:37:12FromDiscord<Phil> Okay is that because you want to do double duty in encoding data into the id or do you just want unguessable numbers?
17:37:37FromDiscord<Chronos [She/Her]> That's the type-
17:37:37FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4EW8
17:37:57FromDiscord<Chronos [She/Her]> In reply to @isofruit "Okay is that because": Double duty?
17:38:32FromDiscord<Chronos [She/Her]> And I'm doing this to prevent the numbers from ever conflicting (since the application may be ran by anyone, all parties not affiliated but the possibility of eachother's servers talking to eachother)
17:38:52FromDiscord<Chronos [She/Her]> It's for a federated chat application idea I have lol
17:39:14FromDiscord<Phil> Ids typically have 1 job:↵Uniquely identify a given record/row in a table.↵If you wish to encode data into them so that you can just fetch the id and just by the id alone know timestamp, systemUniqueValue etc. that means they now have 2 jobs:↵Identify a row and contain data
17:39:32FromDiscord<Phil> (edit) "Ids typically have 1 job:↵Uniquely identify a given record/row in a table.↵If you wish to encode data into them so that you can just fetch the id and just by the id alone know timestamp, systemUniqueValue etc. that means they now have 2 jobs:↵Identify a row and contain data ... " added "that you can parse out of them"
17:40:50FromDiscord<Chronos [She/Her]> Ah, yeah, it needs to be unique at all costs and contain the timestamp info
17:41:07FromDiscord<Phil> Okay so you must be able to parse the timestamp back out of it
17:41:19FromDiscord<Phil> Do you need to get the other things out of it as well?
17:41:37FromDiscord<odexine> if you wanted "unique at all costs" you'd prolly just ask for a kilobyte of random value
17:42:00FromDiscord<Phil> Actually you know what, make an extra timestamp column and generate a UUID based on the 5 numbers provided by you here
17:42:20FromDiscord<odexine> tbh 256 or 512 bytes of random value is also already stupid large
17:42:49FromDiscord<Chronos [She/Her]> In reply to @isofruit "Okay so you must": Only the timestamp
17:42:53FromDiscord<odexine> people are "fine" with UUIDs and those are only 128 bytes minus a little overhead IIRC
17:43:06FromDiscord<odexine> oh bit
17:43:10FromDiscord<odexine> im forgetful
17:43:27FromDiscord<odexine> i've mangled my mind with a lot of embedded systems research today
17:43:33FromDiscord<Phil> double actually, you can just concat the numbers as strings to one another and try to store that as number or sth
17:43:36FromDiscord<odexine> lots of switching between bits and bytes
17:44:07FromDiscord<Chronos [She/Her]> I could probably do that tbf
17:44:35FromDiscord<Phil> Don't try to be clever and have your id-column also act as your storage for timestamp.↵Have whatever you want stored as individual columns and build your id together from your building blocks solely for the purposes of being an id
17:45:07FromDiscord<Phil> Having a dedicated timestamp column also means you're not screwed later when you want to sort stuff by timestamp or select stuff based on timestamp on your postgres
17:46:59FromDiscord<Chronos [She/Her]> Fair enough, but that still means I have to store 104 bits-
17:47:14FromDiscord<Chronos [She/Her]> (13 bytes)
17:47:39FromDiscord<Chronos [She/Her]> Which still doesn't solve the issue of not being able to store it in one type- Maybe as a string but
17:48:04*Figworm joined #nim
17:48:06FromDiscord<Phil> norm doesn't do string ids, another one of those "I don't get why" things
17:48:14FromDiscord<Chronos [She/Her]> Hm
17:48:22FromDiscord<Phil> Actually you could work around that
17:48:35FromDiscord<Phil> Let norm have its id field and just have an additional uuid: string field with a PK constraint
17:48:38FromDiscord<Chronos [She/Her]> If there was a UUID type built in to Norm, that'd be perfect
17:48:49FromDiscord<Chronos [She/Her]> Unless it does and I'm just blind
17:48:54FromDiscord<Phil> It doesn't
17:49:19*Figworm quit (Client Quit)
17:50:40FromDiscord<Chronos [She/Her]> Yeah thought so
17:50:43FromDiscord<Chronos [She/Her]> Sigh...
17:51:21FromDiscord<odexine> would norm allow it to be an array of 2 uint64s
17:52:26FromDiscord<Phil> Norm has hard-coded requirement for an id of int64.↵Or rather it has the hard-coded requirement that such a column must exist and that it be unique so norm can fetch records from the db based on that id
17:53:04FromDiscord<Phil> That doesn't mean you can't have your own UUID field and use raw-selects, it's fairly ugly though. Might still work well enough
17:53:10FromDiscord<Chronos [She/Her]> Does it allow for it to be an int128 instead...?
17:53:19FromDiscord<Chronos [She/Her]> If I could get that to work, that'd be nice
17:54:08FromDiscord<Phil> Nope.↵You inherit from Model, Model forces an int64
17:54:12FromDiscord<Chronos [She/Her]> Damn it
17:54:19FromDiscord<Chronos [She/Her]> How about if I just, fork Norm
17:54:32FromDiscord<odexine> isnt it kinda recommended to have separate "db id" and "external id"s anyway
17:54:39FromDiscord<Chronos [She/Her]> Is it?
17:55:05FromDiscord<odexine> yes as a random pk id has database index performance implications
17:55:07FromDiscord<Phil> beats me. I haven't encountered that pattern yet, but I have not yet worked on a codebase with a db that uses techniques younger than 10 years old
17:55:35FromDiscord<Phil> At least not a relational db
17:55:43FromDiscord<odexine> iirc this is esp true for neosql dbs where theres a large distributed-capable focus
17:55:55FromDiscord<odexine> not sure tho
17:56:00FromDiscord<Phil> And my own work in my spare time never hits those usecases
17:56:06FromDiscord<odexine> yeah
17:56:16FromDiscord<odexine> its not required for us mini-scale programmers
17:56:22FromDiscord<odexine> but its a recommendation
17:58:49FromDiscord<Phil> I mean the only reason forces the id field is so that it can ensure that the field "id" exists
17:59:00FromDiscord<Phil> And concepts at the time were far less promoted afaik
17:59:13FromDiscord<Phil> (edit) "I mean the only reason ... forces" added "norm"
17:59:41FromDiscord<Phil> In reply to @chronos.vitaqua "How about if I": Actually you could still fork norm. Or rather work on a PR that removes the need for inheritance and just adds a `Model` Concept instead
18:00:41FromDiscord<Chronos [She/Her]> I could try that? But also I'm not confident enough on my Nim knowledge to do that lol, esp without an actual editor (I'm using nano on my phone)
18:01:01FromDiscord<Phil> Would likely need some tinkering so that `createTables` can know inherently what database-column-type to use
18:01:35FromDiscord<Phil> OOOOOOORRRRR models could have a pragma annotation in which the user must provide the name of the id-column
18:01:57FromDiscord<Phil> Or just have a pk pragma, even easier
18:02:02FromDiscord<Phil> (edit) "pragma," => "pragma on the specific field,"
18:03:57FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EWh
18:04:54FromDiscord<Chronos [She/Her]> In reply to @isofruit "OOOOOOORRRRR models could have": So a `normId` pragma?
18:05:23FromDiscord<Chronos [She/Her]> sent a code paste, see https://play.nim-lang.org/#ix=4EWj
18:05:58FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EWk
18:06:05FromDiscord<System64 ~ Flandre Scarlet> In reply to @chronos.vitaqua "Can you give a": No error, no stacktrace↵it justs quit
18:06:11FromDiscord<Phil> In reply to @chronos.vitaqua "So a `normId` pragma?": just pk should suffice.↵Supporting partial pk's though would be absolute hell
18:06:18FromDiscord<Phil> (edit) "partial" => "composite"
18:06:19FromDiscord<Chronos [She/Her]> Fair
18:06:49FromDiscord<Chronos [She/Her]> In reply to @sys64 "No error, no stacktrace": Hm... Not sure then
18:07:12FromDiscord<System64 ~ Flandre Scarlet> In reply to @chronos.vitaqua "Hm... Not sure then": I can't even catch the error
18:07:42FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4EWl
18:07:52FromDiscord<Phil> (edit) "https://play.nim-lang.org/#ix=4EWl" => "https://play.nim-lang.org/#ix=4EWm"
18:08:37FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "type of instanceMan plx.": instance Man shouldn't be nil but will check after the dinner
18:08:59FromDiscord<Phil> In reply to @sys64 "instance Man shouldn't be": But the fields you access on it might
18:09:53FromDiscord<System64 ~ Flandre Scarlet> It crashes once I do instance List.del()
18:12:26FromDiscord<System64 ~ Flandre Scarlet> Is it related to multithreading?
18:18:43FromDiscord<Phil> one sec, how does that even work? I assume instanceList is a seq, right?
18:18:59FromDiscord<Phil> This is the kind of reason why I ask to see types, I hate guessing
18:19:02FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "one sec, how does": It's a table
18:19:10FromDiscord<Phil> ... why call it list
18:19:33FromDiscord<System64 ~ Flandre Scarlet> With string as key, and GameInstance as value
18:19:43FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "... why call it": No idea
18:22:42FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4EWo
18:22:59FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "This shouldn't be possible": It's a ref type
18:27:48FromDiscord<Phil> In reply to @sys64 "No error, no stacktrace": I assume you know for 100% certain based on echo's that you never get past the line ` instanceMan.instanceList.del(code)`?
18:29:18FromDiscord<System64 ~ Flandre Scarlet> Yeah, I tried with echo, I don't get past this
18:30:52FromDiscord<Phil> sent a code paste, see https://play.nim-lang.org/#ix=4EWp
18:32:27FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EWq
18:32:28FromDiscord<Phil> sent a code paste, see https://paste.rs/FlOoi
18:32:29FromDiscord<Phil> yes
18:32:46FromDiscord<Chronos [She/Her]> How hard would it be to make my own ORM hm...
18:32:48FromDiscord<System64 ~ Flandre Scarlet> GameInstance is a ref type
18:34:43FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EWs
18:38:52FromDiscord<Phil> sent a long message, see http://ix.io/4EWx
18:40:25FromDiscord<Phil> (edit) "http://ix.io/4EWx" => "http://ix.io/4EWy"
18:42:11FromDiscord<System64 ~ Flandre Scarlet> ``instanceMan.instanceList[code] = nil`` doing that makes the software crashing even faster
18:42:20FromDiscord<Chronos [She/Her]> In reply to @isofruit "For the most part": Makes sense
18:43:04FromDiscord<Phil> In reply to @sys64 "``instanceMan.instanceList[code] = nil`` doing": I mean, that is you assigning nil to your table, that's on you at that point 😛
18:48:25FromDiscord<System64 ~ Flandre Scarlet> In reply to @isofruit "I mean, that is": Now the problem is I don't have any stacktrace so the problem is hard to diagnose
18:49:30FromDiscord<Phil> With today having been exhausting and me not having any source code, I'll leave it at that, maybe beef can figure this one out ^^
18:49:54FromDiscord<Phil> (edit) "any" => "half the"
18:50:58FromDiscord<System64 ~ Flandre Scarlet> This is the entire instangeManager https://media.discordapp.net/attachments/371759389889003532/1146517433318916266/instanceManager.nim
18:52:37FromDiscord<System64 ~ Flandre Scarlet> Humm is it possible to make a thread quit without quitting the entire software?
18:56:36advesperacitmake the thread return, the quit() proc will end everything
18:57:25FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EWG
18:58:59advesperacityes, the defer wraps the rest of the code in a try with your defer clause being the finally section
18:59:33FromDiscord<System64 ~ Flandre Scarlet> Oh alright↵Well, it still crash :/
19:01:15FromDiscord<odexine> sent a code paste, see https://paste.rs/G9c08
19:01:34FromDiscord<griffith1deadly> sent a code paste, see https://play.nim-lang.org/#ix=4EWK
19:01:58FromDiscord<odexine> game.game.gameCode
19:02:02FromDiscord<odexine> add more games
19:02:56FromDiscord<sOkam! 🫐> Is there any difference between `nim c --mm:none thefile.nim` and `nim c --mm:none --noMain thefile.nim`?↵Is the nim's `NimMain()` function required for something other than the GC? 🤔
19:07:18*ntat quit (Quit: leaving)
19:08:56FromDiscord<Phil> https://forum.nim-lang.org/t/10446
19:09:05FromDiscord<Phil> For you sOkam!
19:09:54FromDiscord<System64 ~ Flandre Scarlet> In reply to @griffith1deadly "game.game ...": Well I should name my variables a better way lol
19:10:27FromDiscord<System64 ~ Flandre Scarlet> In reply to @odexine "why is the defer": I want to send a message to a channel so the GameInstance gets dereferenced from a list
19:11:31FromDiscord<sOkam! 🫐> In reply to @isofruit "https://forum.nim-lang.org/t/10446": your search skills are brilliant. better than mine. ty 🙏
19:11:49FromDiscord<Phil> Nah, just knew the thread as I had read it earlier
19:11:51FromDiscord<sOkam! 🫐> that explains a lot
19:11:54FromDiscord<Phil> earlier being a couple days ago
19:11:55FromDiscord<sOkam! 🫐> i see
19:12:47FromDiscord<sOkam! 🫐> very very useful info, ngl. ty a lot
19:16:10FromDiscord<sOkam! 🫐> @pmunch do you know any tool where you can give it a C symbol name, like `glfwInit()`, a .c/.h file to find its implementation, and it can give you back a dependency graph of all other symbols/functions it depends on?↵↵Asking because you made futhark, and I know you used clang for it, but since futhark doesn't do this I wondered if you know how to do it
19:18:00FromDiscord<sOkam! 🫐> The idea is to have some sort of tool where I can slowly pull pieces of a huge C project, to start converting them into nim, but the project is gigantic so I would need to do it slowly to not break everything (and probably will always be a mixed project, never fully pure nim)↵Can you think of anything?
19:18:43FromDiscord<sOkam! 🫐> (edit) "The idea is to have ... some" added "(or create)"
19:19:16FromDiscord<System64 ~ Flandre Scarlet> In reply to @sys64 "I want to send": I think I will have to find another approach
19:19:50FromDiscord<Chronos [She/Her]> In reply to @heysokam "<@392962235737047041> do you know": I'd imagine clang would have a tool for this, no? Or someone using it to of made a tool at l
19:19:53FromDiscord<Chronos [She/Her]> Least
19:46:36FromDiscord<pmunch> In reply to @heysokam "<@392962235737047041> do you know": Hmm, not sure. The definition (.h) alone wouldn't tell it enough to know what else it needs, but you would be able to get the other imports from that file. Of course that is probably enough since including that .h file would pull in the other things as well. Opir sorta does this already. That being said why dont you just wrap the whole thing with Futhark and then you can replace
19:46:59FromDiscord<nnsee> In reply to @heysokam "<@392962235737047041> do you know": i remember doxygen being able to do that
19:47:04FromDiscord<nnsee> i think
19:53:05FromDiscord<sOkam! 🫐> In reply to @pmunch "Hmm, not sure. The": Don't know if you recall, but I asked about that "why dont you just ..." part like 3-4 different times in here, and I still have no single clue how to do it↵It's a beast of a project, its taking me 1+year to even be able to use it from nim at all↵I could wrap glfw no problem, and used futhark to provide an alternative raw version for the wgpu bindings↵But this project is just something
19:56:20FromDiscord<sOkam! 🫐> I need to reimplement the `main()` function for C, and that function is platform-dependent and calling some internal sdl implementation↵So to be able to call the engine code from Nim with futhark at all, first I need to reimplement that entire file and/or wrap it. Which is the root of that conversion process question, basically
19:57:13FromDiscord<pmunch> Well, if you wrap all the includes in your project with Futhark then rewriting the main should be pretty simple
19:57:50FromDiscord<sOkam! 🫐> what do you call with futhark to do that? none of the internal unix_main.c and win32_main.c functions are exported anywhere
19:58:20FromDiscord<sOkam! 🫐> they are all static functions only used in that file, which is the root of the problem to solve
20:01:22FromDiscord<System64 ~ Flandre Scarlet> Does generated C code use the preprocessor?
20:14:57FromDiscord<juan_carlos> In reply to @sys64 "Does generated C code": Ye
20:15:32FromDiscord<System64 ~ Flandre Scarlet> Oh alright, thanks
20:15:46FromDiscord<System64 ~ Flandre Scarlet> I still need to understand this step better for my school work
20:19:49FromDiscord<Chronos [She/Her]> Sigh, wish there was a barebones android app someone made to load a .so file
20:20:07FromDiscord<Chronos [She/Her]> So I could work on a small game with Raylib on my phone
20:49:34FromDiscord<Elegantbeef> can't termux do gui?
20:49:54FromDiscord<ahungry> sent a long message, see http://ix.io/4EX5
20:52:28FromDiscord<ahungry> of course, drawing vs printing text...big diff 😆
20:53:46FromDiscord<ahungry> https://github.com/raysan5/raylib/wiki/Working-for-Android-(on-Linux) - nevermind, solved space - neat - I should give raylib a try sometime
20:57:42FromDiscord<Chronos [She/Her]> In reply to @Elegantbeef "can't termux do gui?": Not in a sane way-
20:58:12FromDiscord<Chronos [She/Her]> In reply to @ahungry "https://github.com/ahungry/janet-android/tree/maste": Ooh
20:58:32FromDiscord<Chronos [She/Her]> Hm but surely it wouldn't be hard to just, modify that to have an entrypoint for my custom so file to use?
21:00:07FromDiscord<jviega> Anyone know off the top of their head how to programatically find the nimble path from within config.nims / nimble file?
21:07:01FromDiscord<sOkam! 🫐> In reply to @chronos.vitaqua "Sigh, wish there was": that sounds like a hella useful project
21:42:54*Jjp137 quit (Quit: Leaving)
21:46:52*Jjp137 joined #nim
21:57:26FromDiscord<.maverk> https://media.discordapp.net/attachments/371759389889003532/1146564358718107658/ee.PNG
21:57:30FromDiscord<.maverk> what does this mean ?
21:57:46FromDiscord<.maverk> why moderated ?
22:04:22FromDiscord<Elegantbeef> Cause your account is new
22:05:18FromDiscord<.maverk> created 25 days ago
22:05:48FromDiscord<.maverk> wtf is this ????? https://media.discordapp.net/attachments/371759389889003532/1146566461805056080/31.PNG
22:06:00FromDiscord<Elegantbeef> It's floats
22:06:01FromDiscord<Elegantbeef> We went over this
22:06:14FromDiscord<.maverk> i thought it is fixed
22:06:24FromDiscord<Elegantbeef> You cannot fix it
22:06:38FromDiscord<Elegantbeef> There is no bug
22:06:46FromDiscord<.maverk> why ? because of IEEX
22:06:52FromDiscord<.maverk> (edit) "IEEX" => "IEEx"
22:07:17FromDiscord<Elegantbeef> Because of floating point
22:07:59FromDiscord<Elegantbeef> https://www.omnicalculator.com/other/floating-point
22:08:07FromDiscord<.maverk> nim doesn't have software for fixing this
22:08:12FromDiscord<Elegantbeef> image.png https://media.discordapp.net/attachments/371759389889003532/1146567069572280470/image.png
22:08:14FromDiscord<Elegantbeef> There is no bug
22:08:26FromDiscord<Elegantbeef> This is what a float is
22:08:49FromDiscord<Elegantbeef> If you mean "I want to format a float to show X significant digits" yes there is formatting for that
22:09:09FromDiscord<Elegantbeef> But `$` is not a formatted string it is the stringified version of a value
22:09:27FromDiscord<jmgomez> sup beef, any idea of what I asked on #internals about how to run a file with multiple specs in testament locally?
22:09:43FromDiscord<Elegantbeef> Nope
22:09:53*def- quit (Quit: -)
22:10:04*def- joined #nim
22:10:45FromDiscord<.maverk> In reply to @Elegantbeef "image.png": so 32 bits are not enough to get the exact floating point ?
22:10:51FromDiscord<.maverk> because 64 works fine
22:11:11FromDiscord<Elegantbeef> Correct
22:11:18FromDiscord<jmgomez> In reply to @Elegantbeef "Nope": 😦
22:15:08*advesperacit quit ()
22:17:35FromDiscord<jmgomez> Crap, is flyx on GH on Discord?
22:17:50FromDiscord<Elegantbeef> Not that I know of
22:18:40FromDiscord<jmgomez> agh he just broke CI with an update to his package and my precious PR towards IC is gonna fail 😦
22:31:27*genr8eofl__ joined #nim
22:31:31FromDiscord<System64 ~ Flandre Scarlet> sent a code paste, see https://play.nim-lang.org/#ix=4EXp
22:33:48*genr8eofl_ quit (Ping timeout: 246 seconds)
23:02:23*jmdaemon joined #nim
23:10:41FromDiscord<ahungry> sent a code paste, see https://play.nim-lang.org/#ix=4EXs
23:39:07*azimut joined #nim